408 ALASKAN GLACIER STUDIES middle of September, when the river was much lower, the rate of advance undoubtedly-exceeded 6$ feet a day. If the advance commenced on September 19, when the river level lowered, the forward movement of more than 825 feet in the 16 days before October 5 was at the rate of over 20 feet a day. Taking the Trrim'rmim of 6$ feet a day, however, the following computation may be made. If the middle of the front of the glacier moved forward at the rate of at least 6^ feet a day when the northern margin was advancing less than half a foot a day (September 19 to October 5, see table, p. 406) then the middle of the glacier may have been moving over 40 feet a day between July 29 and August 6 when the northern margin had an average velocity of nearly 3 feet a day. Repeating the same calculation on the basis of movement at the rate of 20 feet a day in mid-glacier up to October 5, it appears that the ice in the middle of Childs Glacier may have been flowing at the rate of at least 180 feet a day between July 29 and August 6. These figures are, of course, mere estimates, but they indicate clearly the order of magnitude of the rates of glacier motion during a spasmodic advance. The ice was too much crevassed to permit the setting up of flags and there were no pinnacles of ice that could be identified from day to day, so that we were unable to make precise measurements of the rate of movement. It seems clear to us, however, that a rate of movement of not less than 80 or 40 feet a day occurred in mid-summer of 1910, in the central part of the glacier; and the rate may well have been several times that amount. The oscillations of the ice front in mid-glacier during the summer of 1910 are shown in the following table: VARIATIONS IN WIDTH or COPPER RIVER IN FRONT OP CHILDS GLACIER WITH OSCILLATIONS or ICE FRONT IN 1909-1910 Date Width of River Advance or Retreat Based on Maps by July 1909 June 3, 1910 Aug. 11, 1910 Aug. 17, 1910 Unknown date Oct. 5, 1910 775-1800 feet 500-650 feet 1065 feet 625-1225 feet Not measured 600-900 feet Beginning to advance Advance 275-650 feet Retreat 415 feet Retreat 160 feet Stationary Advance 325 l feet Railway Engineers « n « ti Nat. Geog. Society Not surveyed Railway Engineers What the glacier might do before the spring of 1911 was of great interest. It might continue to advance; or, as the dimim'sln'Tig rate of advance on the northern margin after August 11 suggested, it was possible that the strongest advance was over. If advance continued, would the glacier move up to and destroy the railway bridge (PI. CXLVm, A), which was only 1571 feet distant from the northern margin (Fig. 59), or would it stop before advancing that far? The bridge cost $1,500,000, and is the key to the new $20,000,000 railway to the copper mines. It is absolutely certain that no corps of engineers living could save the bridge and railway if the glacier should advance that far. The railway was less than a mile from the middle of the glacier, which might easily have advanced this distance between May and October, i Plus unmeasured retreat after August 17.